📚 Conference Level Research
Researching methods on giving recommendations for anime, based on its contents such as title and genres. The feature of the content will be extracted using BERT Method and compared using Cosine Similarity.
In this project, I was responsible for conducting in-depth research on the model used, as well as writing and drafting our scientific paper. We developed an anime recommendation system that utilizes BERT to contextually understand anime descriptions, and then calculates similarity using Cosine Similarity.
I was the main presenter at the conference where we published the paper - representing the team to present the research and technical results of the model in front of academics and industry practitioners.
Utilized BERT for contextual understanding of anime descriptions
Implemented cosine similarity for accurate recommendation matching